Endogenous GLP-1 mediates postprandial reductions in activation in central reward and satiety areas in patients with type 2 diabetes

Abstract

Aims/hypothesis: The central nervous system (CNS) is a major player in the regulation of food intake. The gut hormone glucagon-like peptide-1 (GLP-1) has been proposed to have an important role in this regulation by relaying information about nutritional status to the CNS. We hypothesised that endogenous GLP-1 has effects on CNS reward and satiety circuits.

Methods: This was a randomised, crossover, placebo-controlled intervention study, performed in a university medical centre in the Netherlands. We included patients with type 2 diabetes and healthy lean control subjects. Individuals were eligible if they were 40-65 years. Inclusion criteria for the healthy lean individuals included a BMI <25 kg/m(2) and normoglycaemia. Inclusion criteria for the patients with type 2 diabetes included BMI >26 kg/m(2), HbA1c levels between 42 and 69 mmol/mol (6.0-8.5%) and treatment for diabetes with only oral glucose-lowering agents. We assessed CNS activation, defined as blood oxygen level dependent (BOLD) signal, in response to food pictures in obese patients with type 2 diabetes (n = 20) and healthy lean individuals (n = 20) using functional magnetic resonance imaging (fMRI). fMRI was performed in the fasted state and after meal intake on two occasions, once during infusion of the GLP-1 receptor antagonist exendin 9-39, which was administered to block actions of endogenous GLP-1, and on the other occasion during saline (placebo) infusion. Participants were blinded for the type of infusion. The order of infusion was determined by block randomisation. The primary outcome was the difference in BOLD signal, i.e. in CNS activation, in predefined regions in the CNS in response to viewing food pictures.

Results: All patients were included in the analyses. Patients with type 2 diabetes showed increased CNS activation in CNS areas involved in the regulation of feeding (insula, amygdala and orbitofrontal cortex) in response to food pictures compared with lean individuals (p ≤ 0.04). Meal intake reduced activation in the insula in response to food pictures in both groups (p ≤ 0.05), but this was more pronounced in patients with type 2 diabetes. Blocking actions of endogenous GLP-1 significantly prevented meal-induced reductions in bilateral insula activation in response to food pictures in patients with type 2 diabetes (p ≤ 0.03).

Conclusions/interpretation: Our findings support the hypothesis that endogenous GLP-1 is involved in postprandial satiating effects in the CNS of obese patients with type 2 diabetes.

Trial registration: ClinicalTrials.gov NCT 01363609. Funding The study was funded in part by a grant from Novo Nordisk.

Trial registration: ClinicalTrials.gov NCT01363609.

Keywords: Food intake; GLP-1; Neuroimaging; Obesity; Type 2 diabetes; fMRI.

Figures

Fig. 1

Study protocol. (a) Study design. Obese patients with type 2 diabetes and healthy lean individuals were studied in a placebo-controlled acute intervention study. The study consisted of two visits (random order): one with a GLP-1 receptor antagonist (exendin 9-39) infusion and one with a saline (placebo) infusion. Infusions started 1 h before the scan and lasted until the end of the visit. During each visit, two fMRI scans were performed: one while fasted and one 30 min after the meal intake. During fMRI, visual-food cues were presented. Blood samples and appetite-related scores on a 10-point Likert scale were taken at fixed time points. T1, structural MRI, T1-weighted sequence. (b) fMRI paradigm. One run comprised six blocks of 21 s each (seven pictures). Within one run, two blocks of each category were presented. Each MRI session included three runs

Fig. 2

Between-group differences on CNS activation in response to viewing food pictures. (a) Axial and (b) coronal slices showing average differences in activation in brain regions where patients with diabetes vs healthy lean individuals had hyperactivation in response to viewing food pictures. The colour scale reflects the T-value of functional activity. Results are presented at the threshold of p < 0.05, FWE corrected on cluster extent. In the graphs, the BOLD signal intensity (effect size [AU]) for each group is plotted as mean and SEM for (c) the right and (d) left insula, (e) right OFC and (f) left amygdala. AU, arbitrary units; FWE, family-wise error; HC, healthy lean controls/individuals; T2DM, type 2 diabetes patients

Fig. 3

Meal intake effects on CNS activation in response to viewing food pictures. Coronal slices showing areas where intake of the meal reduced activation in response to viewing food pictures 30 min after intake in (a) healthy lean individuals and (b) obese patients with diabetes. The colour scale reflects the T-value of functional activity. Results are presented at the threshold of p < 0.05, FWE corrected on cluster extent. In the graphs, the BOLD signal intensity (effect size [AU]) mean and SEM is plotted for healthy lean individuals in (c) the right insula and for patients with diabetes in (d) the right and (e) left insula. AU, arbitrary units; FWE, family-wise error; HC, healthy lean controls/individuals; T2DM, type 2 diabetes patients

Fig. 4

Effects of GLP-1 receptor blockade on CNS responses. Axial and coronal slices showing average differences in activation in brain regions where blockade of endogenous GLP-1 effects with exendin 9-39 prevented reducing effects of meal intake on activation to viewing food pictures in (a) healthy lean individuals (right insula p = 0.08) and (b) patients with type 2 diabetes (bilateral insula p < 0.05). The colour scale reflects the T-value of functional activity. In the graphs, the BOLD signal intensity (effect size [AU]) mean and SEM for healthy lean individuals in (c) the right insula and in patients with diabetes in (d) the right and (e) left insula. The effect of exendin 9-39 in patients with diabetes in response to viewing high-energy food pictures is shown for (f) the right OFC (p = 0.04) and (g) left caudate nucleus (p = 0.06) and left insula (p = 0.08). In the graphs, the signal intensity is plotted for (h) the right OFC, (i) left caudate nucleus and (j) left insula. AU, arbitrary units; ex9-39, exendin 9-39; high en., high-energy food pictures; plac, placebo

Fig. 5

Glucose and plasma hormone levels. Levels of (a) glucose, (b) total GLP-1, (c) insulin and (d) glucagon during placebo (black) and exendin 9-39 (white) infusion in healthy lean individuals (circles) and obese patients with diabetes (squares). Data are mean ± SEM. Glucose levels were higher in diabetic patients vs healthy lean individuals (p < 0.001). Exendin 9-39 administration had no effect on glucose levels in healthy lean individuals (p = 0.4), but increased glucose levels in diabetic patients (p = 0.001). GLP-1 levels were higher during exendin 9-39 vs placebo administration (healthy, lean p = 0.04; diabetes p = 0.002). Insulin levels did not differ between groups nor between infusions in both groups (p ≥ 0.09). Glucagon levels were significantly higher in diabetic vs healthy lean individuals, and in both groups during exendin 9-39 vs placebo administration (p ≤ 0.004)